Détection MV-TFAC dynamique dans un fouillis de mer hétérogène

We propose a ML-CFAR detector which models different intensity areas in the reference window. Thus, it enables to improve classical CFAR detectors in clutter-edge environments and we present its application to spatially correlated sea clutter. An extension which models the texture component of the compound Gaussian clutter vector as the sum of sinusoids is also proposed to improve performances.